凸显尖锐特征的点-线-面递进式曲面重建
|
摘要
针对现有曲面重建算法不能很好地重建出点云模型尖锐特征的缺陷,提出了一种凸显点云尖锐特征的点-线-面递进式曲面重建算法。首先,根据近邻点的欧氏距离、法向偏差和曲面变分,采用主成分分析算法和k-近邻点迭代加权法获取点云准确法向;接着,依据特征点位于多个平面交线上的原则,利用法向聚类和平面拟合从候选特征点中筛选特征点;然后,依据特征点生长方向和主方向的相互关系重建特征线,并按照最小二乘原理采用矩阵法修复角点;最后,以特征线为约束重建尖锐特征点云曲面。实验结果表明:本文算法计算的点云准确法向与理论法向偏差接近于0,特征重建效果优于其他算法,算法耗时短且与点云数量呈线性关系。算法不仅能够准确计算尖锐特征区域的点云法向,还能准确提取出点云模型的特征点并凸显模型的尖锐特征。
To overcome the shortcomings of reconstructing sharp features of the point cloud model in current surface reconstruction algorithms,agradual point-line-surface reconstruction method emphasizing the sharp features was proposed.First,the accurate vectors of the point cloud were calculated using principal component analysis along with k-neighbors iterative weighting according to spatial distance,normal deviation,and surface variation.Second,based on the principle that feature points were present on the intersection area of several planes,the feature points were screened out by vector clustering and plane fitting from the shortlisted feature points.Next,feature lines were reconstructed according to the mutual relation between growing direction and principal direction of the feature points.Meanwhile,corner points were optimized by matrix method on the basis of the least squarestheory.Finally,the surface was reconstructed using the feature lines constraints.Experimental results indicate that the deviation between accurate vectors and theoretical vectors is close to 0.The effect of the proposed algorithm is more superior to the Poisson and MPU algorithms;moreover,the effect is equivalent to the algorithm from reference[4].Furthermore,there exists a linear relationship between the time and point number,and the algorithm consumes less time.The proposed algorithm can accurately estimate the vectors of the point cloud with sharp features;simultaneously,this algorithm can precisely extract feature points from the point cloud models while emphasizing the sharp features of the models.
To overcome the shortcomings of reconstructing sharp features of the point cloud model in current surface reconstruction algorithms,agradual point-line-surface reconstruction method emphasizing the sharp features was proposed.First,the accurate vectors of the point cloud were calculated using principal component analysis along with k-neighbors iterative weighting according to spatial distance,normal deviation,and surface variation.Second,based on the principle that feature points were present on the intersection area of several planes,the feature points were screened out by vector clustering and plane fitting from the shortlisted feature points.Next,feature lines were reconstructed according to the mutual relation between growing direction and principal direction of the feature points.Meanwhile,corner points were optimized by matrix method on the basis of the least squarestheory.Finally,the surface was reconstructed using the feature lines constraints.Experimental results indicate that the deviation between accurate vectors and theoretical vectors is close to 0.The effect of the proposed algorithm is more superior to the Poisson and MPU algorithms;moreover,the effect is equivalent to the algorithm from reference[4].Furthermore,there exists a linear relationship between the time and point number,and the algorithm consumes less time.The proposed algorithm can accurately estimate the vectors of the point cloud with sharp features;simultaneously,this algorithm can precisely extract feature points from the point cloud models while emphasizing the sharp features of the models.
引文
[1] OHTAKE Y,BELYAEV A,ALEXA M,et al..Multi-level partition of unity implicits[J].ACM Transactions on Graphics,2003,22(3):463-470.
[2]李桂清,马维银,鲍虎军.带尖锐特征的Loop细分曲面拟合系统[J].计算机辅助设计与图形学学报,2005,17(6):1179-1185.LI G Q,MA W Y,BAO H J.Fitting system using Loop subdivision surfaces with sharp features[J].Journal of Computer-Aided Design&Computer Graphics,2005,17(6):1179-1185.(in Chinese)
[3] DEY T K,WANG L.Voronoi-based feature curves extraction for sampled singular surfaces[J].Computer and Graphics,2013,37(6):659-668.
[4]李国俊.基于Delaunay细化的散乱点云曲面重建研究[D].郑州:信息工程大学,2015.LI G J.Research on surface reconstruction based on Delaunay refinement from scattered point clouds[D].Zhengzhou:Information Engineering University,2015.(in Chinese)
[5] GUMHOLD S,WAN X,MACLEOD R.Feature extraction from point clouds[C].Proceedings of the 10th International Meshing Roundtable,Berlin:Springer Press,2003:293-305.
[6] PAULY M,KEISER R,GROSS M.Multi-scale feature extraction on point-sampled surfaces[J].Computer Graphics Forum, 2003, 22(3):281-289.
[7] DANIELS J,HA L K,OCHOTTA T,et al..Robust smooth feature extraction from point clouds[J].IEEE International Conference on Shape Modeling and Applications 2007(SMI 07),2007:123-136.
[8]MRIGOT Q,OVSJANIKOV M,GUIBAS L J.Voronoi-Based curvature and feature estimation from point clouds[J].IEEE Trans on Vis and Comput Graph,2011,17(6):743-756.
[9]WEBER C,HAHMANN S,HAGEN H.Sharp feature detection in point clouds[C].IEEE International Conference on Shape Modeling and Application,Boston,2010.
[10]吾守尔·斯拉木,曹巨明.一种新的散乱点云尖锐特征提取方法[J].西安交通大学学报,2012,46(12):1-5,73.WUSHOUR S,CAO J M.An extraction algorithm for sharp feature points from point clouds[J].Journal of Xi′an Jiaotong University,2012,46(12):1-5,73.(in Chinese)
[11]王小超,刘秀平,李宝军,等.基于局部重建的点云特征点提取[J].计算机辅助设计与图形学学报,2013,25(5):659-665.WANG X CH,LIU X P,LI B J,et al..Feature detection on point cloud via local reconstruction[J].Journal of Computer-Aided Design&Computer Graphics,2013,25(5):659-665.(in Chinese)
[12]李明磊,李广云,王力,等.3D Hough Transform在激光点云特征提取中的应用[J].测绘通报,2015,(2):29-33.LI M L,LI G Y,WANG L,et al..Automatic feature detecting from point clouds using 3D Hough Transform[J].Bulletin of Surveying and Mapping,2015,(2):29-33.(in Chinese)
[13]梁栋,王红平,刘修国,等.基于平面基元组的建筑物场景点云自动配准方法[J].武汉大学学报(信息科学版),2016,41(12):1613-1618.LIANG D,WANG H P,LIU X G,et al..Automatic registration of building′s point clouds based on planar primitive groups[J].Geomatics and Information Science of Wuhan University,2016,41(12):1613-1618.(in Chinese)
[14]李明磊,李广云,王力,等.采用八叉树体素生长的点云平面提取[J].光学精密工程,2018,26(1):172-183.LI M L,LI G Y,WANG L,et al.Planar feature extraction from unorganized point clouds using octree voxel-based region growing[J].Opt.Precision Eng.,2018,26(1):172-183.(in Chinese)
[15]李国俊,李宗春,孙元超,等.利用Delaunay细分进行噪声点云曲面重建[J].武汉大学学报(信息科学版),2017,42(1):123-129.LI G J,LI Z CH,SUN Y CH,et al..Using Delaunay refinement to reconstruct surface from point clouds[J].Geomatics and Information Science of Wuhan University,2017,42(1):123-129.(in Chinese)
[16]何华,李宗春,闫荣鑫,等.引入曲面变分实现点云法矢一致性调整[J].测绘学报,2018,47(2):275-280.HE H,LI Z CH,YAN R X,et al..On the consistent normal vector adjustment of point cloud using surface variation[J].Acta Geodaetica et Cartographica Sinica,2018,47(2):275-280.(in Chinese)
[17]袁小翠,吴禄慎,陈华伟.尖锐特征曲面散乱点云法向估计[J].光学精密工程,2016,24(10):2581-2588.YUAN X C,WU L SH,CHEN H W.Normal estimation of scattered point cloud with sharp feature[J].Opt.Precision Eng.,2016,24(10):2581-2588.(in Chinese)
[18] PAULY M,GROSS M,KOBBELT L P.Efficient simplification of point-sampled surfaces[J].Visualization,2002,1(4):163-170.
[19]邹冬.点云模型的尖锐特征提取与分片分析[D].南京:南京师范大学,2012.ZHOU D.Sharp feature extraction and segmentation analysis for point cloud models[D].Nanjing:Nanjing Normal University,2012.(in Chinese)
[20] HAN L,BANCROFT J C.Nearest approaches to multiple lines in n-dimensional space[R].Calgary:CREWES,2010.
[2]李桂清,马维银,鲍虎军.带尖锐特征的Loop细分曲面拟合系统[J].计算机辅助设计与图形学学报,2005,17(6):1179-1185.LI G Q,MA W Y,BAO H J.Fitting system using Loop subdivision surfaces with sharp features[J].Journal of Computer-Aided Design&Computer Graphics,2005,17(6):1179-1185.(in Chinese)
[3] DEY T K,WANG L.Voronoi-based feature curves extraction for sampled singular surfaces[J].Computer and Graphics,2013,37(6):659-668.
[4]李国俊.基于Delaunay细化的散乱点云曲面重建研究[D].郑州:信息工程大学,2015.LI G J.Research on surface reconstruction based on Delaunay refinement from scattered point clouds[D].Zhengzhou:Information Engineering University,2015.(in Chinese)
[5] GUMHOLD S,WAN X,MACLEOD R.Feature extraction from point clouds[C].Proceedings of the 10th International Meshing Roundtable,Berlin:Springer Press,2003:293-305.
[6] PAULY M,KEISER R,GROSS M.Multi-scale feature extraction on point-sampled surfaces[J].Computer Graphics Forum, 2003, 22(3):281-289.
[7] DANIELS J,HA L K,OCHOTTA T,et al..Robust smooth feature extraction from point clouds[J].IEEE International Conference on Shape Modeling and Applications 2007(SMI 07),2007:123-136.
[8]MRIGOT Q,OVSJANIKOV M,GUIBAS L J.Voronoi-Based curvature and feature estimation from point clouds[J].IEEE Trans on Vis and Comput Graph,2011,17(6):743-756.
[9]WEBER C,HAHMANN S,HAGEN H.Sharp feature detection in point clouds[C].IEEE International Conference on Shape Modeling and Application,Boston,2010.
[10]吾守尔·斯拉木,曹巨明.一种新的散乱点云尖锐特征提取方法[J].西安交通大学学报,2012,46(12):1-5,73.WUSHOUR S,CAO J M.An extraction algorithm for sharp feature points from point clouds[J].Journal of Xi′an Jiaotong University,2012,46(12):1-5,73.(in Chinese)
[11]王小超,刘秀平,李宝军,等.基于局部重建的点云特征点提取[J].计算机辅助设计与图形学学报,2013,25(5):659-665.WANG X CH,LIU X P,LI B J,et al..Feature detection on point cloud via local reconstruction[J].Journal of Computer-Aided Design&Computer Graphics,2013,25(5):659-665.(in Chinese)
[12]李明磊,李广云,王力,等.3D Hough Transform在激光点云特征提取中的应用[J].测绘通报,2015,(2):29-33.LI M L,LI G Y,WANG L,et al..Automatic feature detecting from point clouds using 3D Hough Transform[J].Bulletin of Surveying and Mapping,2015,(2):29-33.(in Chinese)
[13]梁栋,王红平,刘修国,等.基于平面基元组的建筑物场景点云自动配准方法[J].武汉大学学报(信息科学版),2016,41(12):1613-1618.LIANG D,WANG H P,LIU X G,et al..Automatic registration of building′s point clouds based on planar primitive groups[J].Geomatics and Information Science of Wuhan University,2016,41(12):1613-1618.(in Chinese)
[14]李明磊,李广云,王力,等.采用八叉树体素生长的点云平面提取[J].光学精密工程,2018,26(1):172-183.LI M L,LI G Y,WANG L,et al.Planar feature extraction from unorganized point clouds using octree voxel-based region growing[J].Opt.Precision Eng.,2018,26(1):172-183.(in Chinese)
[15]李国俊,李宗春,孙元超,等.利用Delaunay细分进行噪声点云曲面重建[J].武汉大学学报(信息科学版),2017,42(1):123-129.LI G J,LI Z CH,SUN Y CH,et al..Using Delaunay refinement to reconstruct surface from point clouds[J].Geomatics and Information Science of Wuhan University,2017,42(1):123-129.(in Chinese)
[16]何华,李宗春,闫荣鑫,等.引入曲面变分实现点云法矢一致性调整[J].测绘学报,2018,47(2):275-280.HE H,LI Z CH,YAN R X,et al..On the consistent normal vector adjustment of point cloud using surface variation[J].Acta Geodaetica et Cartographica Sinica,2018,47(2):275-280.(in Chinese)
[17]袁小翠,吴禄慎,陈华伟.尖锐特征曲面散乱点云法向估计[J].光学精密工程,2016,24(10):2581-2588.YUAN X C,WU L SH,CHEN H W.Normal estimation of scattered point cloud with sharp feature[J].Opt.Precision Eng.,2016,24(10):2581-2588.(in Chinese)
[18] PAULY M,GROSS M,KOBBELT L P.Efficient simplification of point-sampled surfaces[J].Visualization,2002,1(4):163-170.
[19]邹冬.点云模型的尖锐特征提取与分片分析[D].南京:南京师范大学,2012.ZHOU D.Sharp feature extraction and segmentation analysis for point cloud models[D].Nanjing:Nanjing Normal University,2012.(in Chinese)
[20] HAN L,BANCROFT J C.Nearest approaches to multiple lines in n-dimensional space[R].Calgary:CREWES,2010.